STAM: A Framework for Spatio-Temporal Affordance Maps

Riccio, Francesco; Capobianco, Roberto; Hanheide, Marc; Nardi, Daniele

doi:10.1007/978-3-319-47605-6_22

Cited by 1 publication

(4 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Then, in order to respect all the constraints and perform the handover, STAM generates a unique map A E by means of φ. In our case, such function reflects the support of the affordance for the handover task as a combination of the two actions [9].…”

Section: Spatio-temporal Affordancesmentioning

confidence: 91%

“…Since the above cited works only represent spatial affordances to improve robot navigation skills, their approach cannot be generalized to encode spatial semantics of different tasks. Instead, we leverage the formalization of spatio-temporal affordances proposed in [9] and we continuously improve the robot policy by exploiting the constraints imposed by spatio-temporal affordances.…”

Section: Related Workmentioning

confidence: 99%

“…More precisely, we establish a connection between the environment (with its operational functionality) and spatiotemporal affordances. According to the definition provided in [9] and slightly modified in Def. 1, a Spatio-Temporal Affordance (STA) is a function that defines areas of the operational environment that afford an action, given a particular state of the world.…”

Section: Spatio-temporal Affordancesmentioning

confidence: 99%

“…While in these works affordances do not change over time and they are statically attached to each object, the state of the environment is highly dynamic and also contains several dynamic entities, such as humans and robots themselves. This inevitably leads to a more complex problem that requires specific representation and learning approaches, such as the Spatio Temporal Affordance Maps (STAM) proposed in [9]. By using STAM, action affordances change according to the state of the environment and the task to be executed.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Learning human-robot handovers through π-STAM: Policy improvement with spatio-temporal affordance maps

Riccio

Capobianco

Nardi

2016

2016 IEEE-RAS 16th International Conference on Humanoid Robots (Humanoids)

Self Cite

View full text Add to dashboard Cite

Human-robot handovers are characterized by high uncertainty and poor structure of the problem that make them difficult tasks. While machine learning methods have shown promising results, their application to problems with large state dimensionality, such as in the case of humanoid robots, is still limited. Additionally, by using these methods and during the interaction with the human operator, no guarantees can be obtained on the correct interpretation of spatial constraints (e.g., from social rules). In this paper, we present Policy Improvement with Spatio-Temporal Affordance Mapsπ-STAM, a novel iterative algorithm to learn spatial affordances and generate robot behaviors. Our goal consists in generating a policy that adapts to the unknown action semantics by using affordances. In this way, while learning to perform a human-robot handover task, we can (1) efficiently generate good policies with few training episodes, and (2) easily encode action semantics and, if available, enforce prior knowledge in it. We experimentally validate our approach both in simulation and on a real NAO robot whose task consists in taking an object from the hands of a human. The obtained results show that our algorithm obtains a good policy while reducing the computational load and time duration of the learning process.

show abstract

Section: Spatio-temporal Affordancesmentioning

confidence: 91%

Section: Related Workmentioning

confidence: 99%

Section: Spatio-temporal Affordancesmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations